【OPS】qwen3-next support triton chunk_gated_delta_rule ops (#4070)

### What this PR does / why we need it?
qwen3-next suppot  triton chunk_gated_delta_rule ops

### co-owners
@OsirisDuan

- vLLM version: v0.11.2

Signed-off-by: shiyuan680 <917935075@qq.com>

tests/e2e/multicard/test_chunk_gated_delta_rule.py

@@ -0,0 +1,33 @@
+import torch
+
+from tests.ut.base import PytestBase
+from vllm_ascend.ops.triton.fla.chunk import chunk_gated_delta_rule
+
+
+class TestChunkGatedDeltaRule(PytestBase):
+
+    def test_triton_fusion_ops(self, mock_moe_env):
+        q = torch.randn(1, 17, 4, 128, dtype=torch.bfloat16).npu()
+        k = torch.randn(1, 17, 4, 128, dtype=torch.bfloat16).npu()
+        v = torch.randn(1, 17, 8, 128, dtype=torch.bfloat16).npu()
+        g = torch.randn(1, 17, 8, dtype=torch.float32).npu()
+        beta = torch.randn(1, 17, 8, dtype=torch.bfloat16).npu()
+        initial_state = torch.randn(3, 8, 128, 128, dtype=torch.bfloat16).npu()
+        q_start_loc = torch.range(0, 3, dtype=torch.int).npu()
+
+        (
+            core_attn_out_non_spec,
+            last_recurrent_state,
+        ) = chunk_gated_delta_rule(q=q,
+                                   k=k,
+                                   v=v,
+                                   g=g,
+                                   beta=beta,
+                                   initial_state=initial_state,
+                                   output_final_state=True,
+                                   cu_seqlens=q_start_loc,
+                                   head_first=False,
+                                   use_qk_l2norm_in_kernel=True)
+
+        assert core_attn_out_non_spec.shape == (1, 17, 8, 128)
+        assert last_recurrent_state.shape == (3, 8, 128, 128)